A new study led by the University of East Anglia reveals that although AI-generated essays are impressively coherent, they lack the personal touch and engagement markers that make student essays more compelling. This research underscores the irreplaceable value of human creativity and critical thinking in education.
In a revealing new study, researchers at the University of East Anglia (UEA), in collaboration with Kevin Jiang, a professor at Jilin University in China, have found that while AI tools like ChatGPT can produce coherent and grammatically correct essays, they fall short of the engaging and personalized work crafted by actual students.
The study analyzed 290 essays — half written by students and half generated by ChatGPT.
The findings, published in the journal Written Communication, highlight a growing concern among educators regarding the impact of AI on academic integrity and the development of critical thinking skills.
AI’s Shortcomings
“Since its public release, ChatGPT has created considerable anxiety among teachers worried that students will use it to write their assignments,” Ken Hyland, a professor in UEA’s School of Education and Lifelong Learning, said in a news release. “The fear is that ChatGPT and other AI writing tools potentially facilitate cheating and may weaken core literacy and critical thinking skills.”
The researchers were particularly focused on “engagement markers,” such as rhetorical questions, personal commentary and direct appeals to the reader.
These elements are crucial for crafting persuasive and interactive essays, and they were notably less frequent in AI-generated texts.
The Human Touch
“We found that the essays written by real students consistently featured a rich array of engagement strategies, making them more interactive and persuasive,” added Hyland. “They were full of rhetorical questions, personal asides and direct appeals to the reader — all techniques that enhance clarity, connection, and produce a strong argument.”
Conversely, AI-generated essays, while linguistically fluent, were often impersonal and lacked a clear stance on the topics discussed.
“They tended to avoid questions and limited personal commentary. Overall, they were less engaging, less persuasive, and there was no strong perspective on a topic,” Hyland added. “This reflects the nature of its training data and statistical learning methods, which prioritize coherence over conversational nuance.”
Educational Implications
Despite these shortcomings, the study does not advocate for abandoning AI tools in educational settings. Instead, it suggests that tools like ChatGPT should serve as educational aids rather than shortcuts.
“When students come to school, college or university, we’re not just teaching them how to write, we’re teaching them how to think — and that’s something no algorithm can replicate,” Hyland added.
Source: University of East Anglia